Drainage, sowing date and variety effects on chickpea grown ......Drainage, sowing date and variety effects on chickpea grown on a Vertisol in Ethiopia Getachew Agegnehu and Woldeyesus

This article was downloaded by: [Mr Getachew Agegnehu]On: 15 February 2012, At: 22:34Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Archives of Agronomy and Soil SciencePublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/gags20

Drainage, sowing date and varietyeffects on chickpea grown on a Vertisolin EthiopiaGetachew Agegnehu a & Woldeyesus Sinebo aa Holetta Agricultural Research Centre, Ethiopian Institute ofAgricultural Research, Addis Ababa, Ethiopia

Available online: 20 Jul 2011

To cite this article: Getachew Agegnehu & Woldeyesus Sinebo (2012): Drainage, sowing date andvariety effects on chickpea grown on a Vertisol in Ethiopia, Archives of Agronomy and Soil Science,58:1, 101-113

To link to this article: http://dx.doi.org/10.1080/03650340.2010.503958

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representationthat the contents will be complete or accurate or up to date. The accuracy of anyinstructions, formulae, and drug doses should be independently verified with primarysources. The publisher shall not be liable for any loss, actions, claims, proceedings,demand, or costs or damages whatsoever or howsoever caused arising directly orindirectly in connection with or arising out of the use of this material.

Drainage, sowing date and variety effects on chickpea grown on a

Vertisol in Ethiopia

Getachew Agegnehu and Woldeyesus Sinebo*

Holetta Agricultural Research Centre, Ethiopian Institute of Agricultural Research,Addis Ababa, Ethiopia

(Received 7 April 2010; final version received 19 June 2010)

A field experiment was conducted for three years at Ginchi in Ethiopia to studythe effects of three drainage methods [broad-bed-and-furrow (BBF), ridge-and-furrow (RF) and flat beds (FB)] arranged as main plots. Sub-plots comprised afactorial combination of four sowing dates (18 and 31 August, and 14 and 28September) and three Desi-type chickpea varieties (Worku, Akaki and a landrace)in a split-plot design with three replications. Improved drainage methods (BBFand RF) increased chickpea seed yield by an average of 45% over the flat seedbed.There was a quadratic relationship between seed yield and sowing date with apeak yield in mid-September. Improved varieties (Worku and Akaki) yielded 15–19% more than the local check. Improved varieties were significantly moreyielding than the landrace variety under the improved drainage system but notunder the flat bed system. Also, improved varieties yielded significantly more thanthe landrace variety in the first three sowing dates when waterlogging was aproblem but not in the last sowing date after which drought stress normally setsin. Sowing of improved chickpea varieties in mid-September using BBF couldmarkedly increase productivity of chickpea on Vertisols in Ethiopia.

Keywords: drainage method; chickpea genotypes; sowing date; Vertisol;waterlogging

Introduction

In Ethiopia, chickpea (Cicer arietinum L.) is an important food legume crop grownmainly on Vertisols in the highlands. About 12.6 million ha of Vertisols are found inEthiopia, making up 10.3% of the total land area and 25% of the cropland area(Wakeel and Astatke 1996). In Ethiopia, despite their high agricultural potential,Vertisols are generally considered problem soils because of poor workability andprolonged waterlogging during the main rainy season (Belayneh 1987; Asamenewet al. 1988; Abebe et al. 1994). To avoid waterlogging stress, crops except tef(Eragrostis tef L. – because it is waterlogging resistant) are sown on Vertisolstowards the end of the main rainy season. Such a late-sown crop mainly utilizesresidual soil moisture for growth and development, often experiencing droughtduring seed filling, resulting in lower yields.

*Corresponding author. Email: [email protected]

Archives of Agronomy and Soil Science

Vol. 58, No. 1, January 2012, 101–113

ISSN 0365-0340 print/ISSN 1476-3567 online

� 2012 Taylor & Francis

http://dx.doi.org/10.1080/03650340.2010.503958

http://www.tandfonline.com

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

In addition to late sowing, farmers cope with waterlogging by constructing aridge-and-furrow (RF) drainage system, whereby parallel ridges of*20 cm high and30 cm wide are made by hand after broadcasting seeds (Erkossa et al. 2006). Thecrops grow on the ridges, allowing the excess water to drain out of the field throughthe furrows. In the RF system, land preparation starts during the April–May smallrains, with occasional ploughing to control weeds throughout the summer untilplanting (Erkossa et al. 2006). To make improvements over the traditional RFdrainage system, an animal-drawn device named the broad-bed maker (BBM) wasintroduced into Ethiopia by the International Livestock Center for Africa (Haqueet al. 1988; Mamo and Mohamed-Saleem 2001; Erkossa et al. 2006). The BBM isused to make broad-bed-and-furrow (BBF) in which 80-cm-wide beds alternate with20-cm-wide furrows. The crops are arranged for growing on the beds with furrowsused to drain excess water.

Chickpea is sensitive to waterlogging. Because of this, in Ethiopia this crop issown starting from about mid to late September, at which time the waterloggingproblem has receded and drought stress is about to set in. The crop is sown with onlyone ploughing before sowing, using the traditional ox-drawn hand plough, followedby another to cover the broadcasted seeds on land that was kept fallow during theheavy rains from June to August. In general, no drainage practices are used forgrowing chickpea. Given the current practices, chickpea productivity on Vertisols isconstrained by severe waterlogging when sown early and by drought when sown late.Improving drainage might enable early sowing, increase the growth period withoutdrought stress, and therefore enhance seed yield.

Early September sowing of chickpea on Vertisol at Debre Zeit increased seedyield by 35% compared with later sowing dates (Eshete 1994). Debre Zeit (east ofAddis Ababa) is largely a continuum of the Rift Valley system having lower altitude,warmer temperatures and less rainfall compared with Ginchi, which is *90 km westof Addis Ababa on the way to the high rainfall belt of western Ethiopia. Being atGinchi or at Debre Zeit, it can be surmised that late season drought stress onchickpea grown on Vertisols can be avoided by sowing even earlier than therecommended early September sowing dates, provided that the problem ofwaterlogging, for example, in August, is tackled. However, the hypothesis thatsowing in August in the presence of improved drainage systems may be moreadvantageous than later sowing has not been supported by research data.Furthermore, whether improved drainage is important for chickpea sown inSeptember has not been determined.

The national chickpea breeding program in Ethiopia has developed severalimproved varieties that are being promoted for large-scale adoption by farmers. Theimproved chickpea varieties differ in important agronomic attributes including seedsize and grain yield. Nonetheless, chickpea landraces continue to be widely grown inlarge parts of Ethiopia. It is often believed that crop landraces display a specificadaptation that may confer yield advantages over modern varieties under abioticstress situations such as drought (Ceccarelli 1994; Sinebo 2002). It may, therefore, beimportant to examine the differential performance of landrace and modern chickpeavarieties when subjected to different levels of management regimes such as sowingdates and drainage methods.

Our objective was to study the main and interaction effects of sowing date,drainage method and variety on growth, seed yield and yield components ofchickpea planted over several years on Vertisol of Ginchi, Ethiopia.

102 G. Agegnehu and W. Sinebo

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

Materials and methods

Experimental site

The trial was conducted for three years (2003–2005) at Ginchi Research Sub-centrein the central highlands of Ethiopia. Ginchi is located at 098020N latitude and388120E longitude at an altitude of 2200 m above sea level. The soil is a pellicVertisol. The soil physical and chemical properties for the experimental site are givenin Table 1. The rainfall is bimodal with long-term average annual rainfall of1095 mm, of which *76% is received from June to September and 24% fromJanuary to May (Table 2).

Land preparation, treatments, design and data collection

Tractor-mounted disc ploughing and disc harrowing was carried out in April as iscommon practice for general land preparation in the research centre. The BBF wasconstructed in June with a broad-bed maker (BBM) pulled by a pair of oxen. BBFwas made in June because the soil was still relatively friable and workable at this timefor using the BBM. Ridge-and-furrow was made by hand immediately before thefirst planting on 18 August for all treatments. At each sowing date, the BBF and RFwere maintained or renewed by hand using hand hoes. Similarly, a hand hoe wasused to disturb the soil before each sowing for the flat seedbed treatments.

The design was a split-plot with three replications. The three drainage methods(broad-beds-and-furrow, ridge-and-furrow and flat beds) were arranged in the mainplots, and factorial combinations of four sowing dates (18 and 31 August, and 14and 28 September) and three Desi-type chickpea varieties (Worku, Akaki andlandrace) were used in the sub-plot. A sub-plot size of 3.6 m 6 4 m was used. Seedswere drilled by hand in 30-cm-wide rows at an interplant spacing of *10 cm andlightly covered with soil. Seed rate of 110 kg ha71 for the improved varieties and100 kg ha71 for the landrace (adjusted for smaller seed size) were used. Fertilizerwas applied at the rate of 18/20 kg N/P ha71. For the BBF, there were four beds, 80-cm-wide, separated by furrows in each plot. The two centre beds, each consisting ofthree rows of plants for a total of six rows per plot, were harvested for yielddetermination. In the case of RF, chickpea was sown in rows by hand on ridges thatwere 30 cm apart. Again the central six rows (six ridges) were harvested for yielddetermination. This means that, in each of the three drainage systems, each sub-plotconsisted of 12 rows, 30 cm apart, and 4 m long. From each sub-plot, the centresix rows were harvested for a net plot size of 6 rows 6 0.3 m wide 6 4 m long ¼

Table 1. Physical and chemical soil characteristics (0–30 cm depth) of the experimental siteat Ginchi, Ethiopia.

Parameter Value Parameter Value

Clay (%) 66.4 P-Olsen (mg kg71) 16.5Silt (%) 22.3 Na (cmol kg71) 0.42Sand (%) 11.3 K (cmol kg71) 2.04pH (1:1 H2O) 6.5 Ca (cmol kg71) 39.1OC (%) 1.8 Mg (cmol kg71) 7.4C/N ratio 10.1 Ca/Mg ratio 5.3:1Total N (%) 0.13 CEC (cmol kg71) 48.5

Archives of Agronomy and Soil Science 103

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

Table2.

Monthly

totalrainfall,monthly

meanmaxim

um

andminim

um

temperature

andmonthly

meansunshinehours

for2003–2005croppingseasons

andthe30-yearaverage.

Year

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sept

Oct

Nov

Dec

Total

Monthly

totalrainfall(m

m)

2003

7.5

48.2

62.4

136

5125

254

280

96

55

51029

2004

53.8

5.6

28.1

156

46

120

211

249

169

11

80

1058

2005

64.7

2.4

97.8

59

135

144

219

218

132

14

60

1092

30-yearmean

21.2

38.2

72.7

90

86

148

221

222

136

38

14

81095

Monthly

meanmaxim

um

temperature

(8C)

2003

25.7

27.6

26.9

26.1

28.4

24.3

20.9

21.1

21.9

24.1

24.7

24.5

24.7

2004

25.9

26.4

27.1

25.7

27.1

23.6

21.5

22.0

22.0

22.8

24.9

25.2

24.5

2005

25.3

28.1

27.2

27.0

25.7

25.1

22.6

22.3

22.0

23.7

24.4

24.8

24.9

30-yearmean

25.2

26.4

26.6

25.9

26.0

23.9

21.4

21.3

22.3

23.6

24.0

24.6

24.3

Monthly

meanminim

um

temperature

(8C)

2003

6.7

89.1

9.8

9.2

9.6

11.2

11.5

10.5

6.8

6.6

5.8

8.7

2004

9.5

8.2

9.4

10.3

10.2

10.9

11.1

11.5

10.7

8.0

6.6

8.1

9.5

2005

3.4

4.5

7.3

8.0

9.4

7.0

8.7

8.8

8.3

4.3

1.9

70.3

5.9

30-yearmean

8.7

8.7

10.7

11.8

11.8

10.5

11.1

11.4

11.1

7.9

6.5

4.8

9.6

104 G. Agegnehu and W. Sinebo

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

7.2 m2. Harvesting was from 11 to 25 January in 2003, 4 to 18 January in 2004 and13 to 27 January in 2005. Agronomic parameters such as plant stand count (plantm72) at complete emergence and at harvest, days to flowering, mature plant height,number of pods and number of seeds per plant, seed weight, total plant biomass andseed yield were recorded.

Statistical analysis

The data were subjected to analysis of variance using the GLM procedure of SASstatistical package version 8.2 (SAS Institute 2001). Data were combined over thethree years and total variability for each trait was quantified using pooled analysis ofvariance over years based on the following model:

Tijklm ¼ mþ Yi þ RðYÞji þDk þ YDik þ RðYDÞjik þ Sl þ YSil þDSkl

þ YDSikl þ Gm þ YGim þDFkm þ SGlm þ YDGikm þ YSGilm

þDSGklm þ YDSGiklm þ eijklm

where Tijklm is total observation, m ¼ grand mean, Yi ¼ effect of the ith year, R(Y)jiis effect of the jth replication within ith year, Dk is effect of the kth drainage method,Sl is effect of the lth sowing date, Gm is effect of the mth variety, YS, YD, YG, SD,SG, DG, YSD, YSG, YDG, SDG and YSDG are the interactions, and R(YD) andeijklm are the variations due to random error for main and sub-plots, respectively.Significance of the year effect was tested against the R(Y) mean square as the errorterm and the D and YD effects tested against the R(YD) mean square as an errorterm. All other effects were tested against the residual. When interaction effects weresignificant, the slice option in the LSMEANS statement of the GLM procedure wasused to determine the significance of simple effects. Means for the main effects wereseparated using the MEANS statement with the LSD option, in particular specifyingthe appropriate error terms for Y and D. Means for the interactions were separatedusing the PDIFF option in the LSMEANS statement of the GLM procedure, inparticular specifying the R(YD) as an appropriate error term for separatingLSMEANS for the YD interaction.

Results

Weather

Rainfall for July and August was greater and for September was less in 2003 than ineither 2004 or 2005 (Table 2), exposing the crop to more waterlogging at the earlygrowth stage and to drought at later growth stages in 2003. When compared with thelong-term average, rainfall in September was near average in 2005, less by 40 mm in2003 and more by 33 mm in 2004, implying a favourable moisture regime in 2004,average conditions in 2005 and stress in 2003.

Year, drainage, sowing date and variety main effects on seed yield

All the main effects [year (Y), drainage (D), sowing date (S) and variety (G)] werehighly significant (p 5 0.001) (Table 3). Mean seed yield was 2519 kg ha71 in 2004,1548 kg ha71 in 2005 and 887 kg ha71 in 2003 (Table 4). On average, drainage

Archives of Agronomy and Soil Science 105

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

increased seed yield by 45% compared with the FB system. There was no significantdifference in seed yield between the BBF and RF drainage systems (Table 4). By andlarge, there was a quadratic response of seed yield to sowing date with a peak atabout mid-September (Figures 1 and 2 and Table 4). Delaying the sowing date from18 August to 14 September increased seed yield but a further delay until 28September resulted in a decrease in seed yield due to a shortage of moisture. Theimproved varieties, Worku and Akaki, yielded significantly more than the landracecheck (Table 4). There was no significant difference in seed yield between the twoimproved varieties.

Interaction effects on seed yield

The two-factor interaction effects of year by drainage method (Y 6 D), year bysowing date (Y 6 S), drainage method by sowing date (D 6 S), drainage methodby variety (D 6 G) and sowing date by variety (S 6 G) were significant (p 5 0.01)for chickpea seed yield (Table 3). The second-order interaction effects of year bydrainage and sowing date (Y 6 D 6 S) and year by drainage and variety(Y 6 D 6 G) were also significant (p 5 0.01) for seed yield (Table 3).

Slicing the Y 6 D 6 S interaction effect by year indicated the D 6 Sinteraction to be significant each year. The pattern of this interaction for individualyears is given in Figure 1. Each year, there was large curvilinear response (R2 of0.97–1) of seed yield to sowing dates in the BBF and flat bed systems. The curvilinearresponse of seed yield to sowing dates in the RF system was weaker with R2

values ranging from 0.64 to 0.83 in individual years. RF yielded more than BBFand FB systems across almost all the sowing dates in 2003 and in the first sowing

Table 3. Analysis of variance for chickpea seed yield and other agronomic traits tested atthree drainage methods, four sowing dates and three varieties, 2003–2005.

Source df SY BY TSW PPP SPP PH DTF SC2 SC1

Year (Y) 2 *** *** *** *** *** *** *** *** nsDrainage (D) 2 *** *** ns *** *** *** ns ns nsY 6 D 4 *** *** * ** * ns ns * *Sowing date (S) 3 *** *** ns *** *** *** *** * ***Y 6 S 6 *** *** ** *** *** ** ns ** ***D 6 S 6 *** *** ns * ns ns ns ns ***Y 6 D 6 S 12 *** *** ns ns * ns ns ns ***Variety (G) 2 *** *** *** * *** ** ns *** nsY 6 G 4 ns ns *** *** *** ** ns *** ***D 6 G 4 *** *** ns *** * ns ns ns nsY 6 D 6 G 8 *** ** ns ns ** ns ns ns nsS 6 G 6 ** ns ns ns ns ns ns ns nsY 6 S 6 G 12 ns ns ns ** ns ns ns ns nsD 6 S 6 G 12 ns * ns ns ns ns ns ns nsY 6 D 6 S 6 G 24 ns ns ns ns ns ns ns ns nsRoot-MSE 301.7 563.3 18.0 15.0 24.6 3.5 9.7 3.0 6.1CV (%) 18.3 18.6 10.0 21.9 26.4 9.2 15.1 10.0 20.1

Note: SY, seed yield; BY, biomass yield; TSW, thousand seed weight; PPP, pods per plant; SPP, seeds perplant; PH, plant height; DTF, days to flowering; SC1, stand count at emergence; SC2, stand count atmaturity. Significant at *p ¼ 0.05, **p ¼ 0.01, ***p ¼ 0.001; ns, not significant.

106 G. Agegnehu and W. Sinebo

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

Table

4.

Yield

andyield

components

ofchickpea

asinfluencedbydrainagemethod,sowingdate

andvarietyonVertisol,2003–2005.

Factor

SY

(kgha7

1)

BY

(kgha7

1)

TSW

(g)

PPP(no.)

SPP

(no.)

PH

(cm)

DTF(d)

SC1

(plantm

72)

SC2

(plantm

72)

Mortality

(%)

Year(Y

)2003

887

1736

190

56

76

37

59

27

18

33

2004

2519

4509

173

86

118

42

68

33

26

21

2005

1548

2819

179

64

85

36

64

32

27

16

LSD

246.0

427.4

4.2

6.0

9.3

2.0

1.5

ns

0.8

Drainage(D

)BBF

1853

3398

182

74

101

39

64

32

26

19

RF

1831

3334

179

71

99

38

65

31

24

22

Flat

1270

2332

181

60

80

36

64

30

22

27

LSD

111.5

242

ns

3.5

7.6

0.9

ns

ns

ns

Sowingdate

(S)

18Aug

1142

2161

176

56

73

36

61

23

13

44

31Aug

1736

3206

182

74

103

39

62

31

24

23

14Sept

1996

3473

182

74

104

39

69

33

27

18

28Sept

1733

3246

183

69

92

38

64

35

29

17

LSD

93.5

174.5

ns

4.6

7.6

1.1

3.0

1.9

0.9

Variety(G

)Worku

1705

3221

218

66

78

39

64

31

24

22

Akaki

1763

3180

195

71

93

37

65

32

23

22

Landrace

1487

2663

129

69

108

38

63

30

21

27

LSD

81.0

151

4.8

4.0

6.6

0.9

2.6

ns

0.8

Note:SY,seed

yield;BY,biomass

yield;TSW,thousandseed

weight;PPP,podsper

plant;SPP,seedsper

plant;PH,plantheight;DTF,daysto

flowering;SC1,standcountat

emergence;SC2,standcountatmaturity.LSD,least

significantdifference.

Archives of Agronomy and Soil Science 107

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

date in all the years. BBF was more yielding than either the RF or flat bed system formid-September sowings in 2004 and 2005. The RF and BBF systems were moreyielding than the flat bed system across almost all the sowing dates in all the threeyears.

Although it is appropriate to present simple effects as above when complex three-way interactions such as Y 6 D 6 S were significant, two-way interactions aremore useful in providing meaningful agronomic insights. Thus, two-way interactionsinvolving sowing dates are presented in Figure 2. When sliced by year, the averageeffect of sowing date in individual years was significant (p 5 0.0001; data notshown). Examination of the Y 6 S interaction indicated no rank change (cross-overs) among the sowing date treatments in the three years (Figure 2a). Hence thisinteraction is due to scale effects. Also, in each of the three years, the maximum yieldwas obtained with the mid-September sowing date, implying a consistency of relativeresponse of sowing dates over the years.

A graphic representation of the pattern of the D 6 S interaction is given inFigure 2b. Differences among sowing dates were highly significant (p 5 0.0001)under each of the three drainage systems. However, the response was large for theFB and BBF systems and very small for the RF drainage system. In all drainagesystems, however, the maximum seed yield occurred around mid-September makingthe general sowing date recommendation robust. The advantage of improved

Figure 1. Interaction effects of year-by-drainage system-by-sowing date on seed yield ofchickpea tested at four sowing dates and three drainage systems using three varieties at Ginchi,Ethiopia. a, 2003; b, 2004; c, 2005. R2 values are for quadratic equations.

108 G. Agegnehu and W. Sinebo

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

drainage was larger for earlier sowings and small for the last sowing. This isunderstandable because further delay in sowing induces low moisture stressesassociated with the cessation of the season’s rainfall.

The sowing date by variety (S 6 G) interaction is displayed graphically in Figure2c. Differences among varieties were significant for all but the last sowing date. Withearly sowing, when waterlogging was high, improved varieties were more yieldingthan the local variety, but with the last sowing date at the end of September when thelow moisture stress had set in, all varieties gave similar seed yields.

When the Y 6 D 6 G interaction was sliced by year, the D 6 G interactionwas significant for each year (data not shown). The pattern of the Y 6 D 6 Ginteraction is given in Figure 3. In 2003, there was no difference among the threevarieties under the BBF and FB systems, but the two improved varieties yieldedmore than the landrace variety under the RF system. In 2004 and 2005, either orboth improved varieties yielded more than the landrace variety under improveddrainage of RF and BBF. By contrast, the landrace variety was as yielding as theimproved varieties under the FB system in 2004 and 2005. There was no significantdifference between the two improved varieties in all but one combination of year anddrainage system (under RF in 2005).

The two-way interactions of drainage with year and variety are given in Figure 4.Slicing the Y 6 D indicated that the interaction was significant each year. The RFdrainage system gave a significantly greater yield than either the FB or BBF systems

Figure 2. The relationship between sowing date and seed yield as affected by (a) year, (b)drainage system and (c) variety in chickpea tested under three drainage methods and threesowing dates in three years using three varieties at Ginchi, Ethiopia.

Archives of Agronomy and Soil Science 109

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

in 2003 (Figure 4a). In 2004 and 2005, the two improved drainage systems gavesignificantly better yields than the FB system. Seed yields were similar for the twoimproved drainage systems in 2005, but BBF resulted in a better yield than RF in2004 (Figure 4a).

The drainage method by variety (D 6 G) interaction is quite interesting (Figure4b). In the improved drainage systems of RF and BBF, improved varieties gavehigher yields than the local cultivar. But under the FB system, there was nosignificant difference in seed yield among the varieties.

Main and interaction effects on other agronomic attributes

Variance analysis results for biomass yield were similar to those of seed yield. Year,sowing date, variety, drainage and Y 6 S interaction significantly affected most ofthe other parameters measured (Table 3). Stand count at maturity was heavilyinfluenced (p 5 0.01) by year, variety and Y 6 G interaction (Table 3). Stand countat harvest was lower (18 plants m72) in 2003 than in 2004 and 2005 (average of 26

Figure 3. Interaction effects of year-by-drainage system-by-variety on seed yield of chickpeatested at four sowing dates, three drainage systems using three varieties in three years atGinchi, Ethiopia. a, 2003; b, 2004; c, 2005.

110 G. Agegnehu and W. Sinebo

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

plants m72). Stand count at harvest increased from 13 plants m72 for the 18August sowing to 29 plants m72 for the 28 September sowing (Table 4). Plantmortality was highest in 2003 and lowest in 2005. Plant mortality was highestwhen sowing was early and the seedbed was flat compared with mortalitypercentages observed in the later sowing dates and the improved BBF and RFdrainage methods (Table 4).

Seed number per plant (SPP) was strongly influenced (p 5 0.01) by year,drainage, sowing date, variety, Y 6 S and Y 6 G interactions (Table 3). Seednumber per plant was highest in 2004 at 118 SPP, compared with 76 SPP in 2003 and85 SPP in 2005 (Table 4). Seed number per plant increased from 73 for the 18 Augustsowing to 104 for the mid-September sowing and decreased to 92 SPP for the lastsowing made on 28 September (Table 4). Seed weight was strongly affected(p 5 0.01) by year, variety and Y 6 G interaction. The landrace variety had thelowest seed weight but the highest SPP (Table 4). Days to flowering was affected byyear and sowing date only. Plants flowered late when sown in mid-September, whichwas also the date at which seed yields were the highest.

Discussion

In Ethiopia, chickpea is a low-input crop grown with minimum tillage mainly onVertisols that are prone to severe seasonal waterlogging. Chickpea is highlysensitive to waterlogging and is, therefore, grown largely with residual moistureafter the main season rain has subsided. This crop faces drought stress, particularlyduring the critical period of seed filling. Hence, chickpea productivity on Vertisolsin Ethiopia is constrained by poor drainage when sown early and by drought whensown late.

The advantage of improved drainage combined with early sowing on Vertisolshas been demonstrated in several other crops (Haque et al. 1988; Abebe et al. 1994;Asamenew et al. 1988; Astatke et al. 2002; Agegnehu et al. 2006) but not onchickpea. This study has revealed that yield improvement on chickpea grown onVertisols can be made by combining the use of optimum sowing dates, improveddrainage systems and varieties with appropriate phenology. However, even with the

Figure 4. Interaction effects of drainage system with (a) year and (b) variety for seed yield ofchickpea tested under three drainage systems and three sowing dates using three varieties inthree years at Ginchi, Ethiopia. RF, ridge and furrow; BBF, broad bed and furrow; FL, flat.

Archives of Agronomy and Soil Science 111

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

best combination of these improved practices, a lethal combination of high rainfallearly in the season and low rainfall towards end of the season, as observed in 2003,can severely affect chickpea growth and reduce seed yield. Low yield in such anadverse year results from poor vegetative growth, sparse stand at harvest and lowernumber of seeds per plant (Table 4).

Belayneh (1987) emphasized the advantage of improved drainage (camber beds)for chickpea yield at Ginchi (Ethiopia), and Bejiga et al. (1994) reported theimportance of optimum sowing dates for obtaining better chickpea seed yield atDebre Zeit (Ethiopia). But the current study is the first to report the interactive effectof drainage methods and sowing date on several chickpea varieties grown on Vertisolin Ethiopia. Of the statistically significant interactions, no cross-overs were observedfor the year by sowing date (Y 6 S) interaction, implying that sowing daterecommendations are consistent over the years (Figure 2a). Similarly, the sowingdate by variety (S 6 G) interaction did not involve significant rank changes,indicating the optimality of mid-September sowing for all the three varieties.Nonetheless, it is important to note from this interaction that there was noadvantage of improved varieties when chickpea was sown in late September (Figure2c). It appears that the improved varieties were more tolerant than the landrace towaterlogging but not to season-end drought.

Although improved drainage systems were better than the flat system in all years,there were rank changes between BBF and RF (Figure 4a). The cross-over was dueto better seed yield from RF than BBF when waterlogging was severe, as in 2003, butlower yield in RF than in BBF when waterlogging was mild, as in 2004 and 2005.The RF system is likely to facilitate drainage more than BBF because the furrows areat shorter intervals in RF (*30 cm) than in BBF (*80 cm). With high drainageconstraints, as was the case with early sowing, RF yielded significantly more thanBBF (Figure 2b). However, with later sowings, because of reduced waterloggingstress, BBF gave higher yields than RF. From this study, it was apparent that sowingearlier than September did not offer any yield advantage with or without drainage,implying that improved drainage did not help in moving optimum planting datefrom September to August.

The highly significant drainage method by variety (D 6 G) interaction for seedyield was due to the greater yield of improved varieties under relatively improveddrainage conditions, but not under flat bed conditions. This result confirms the oftenclaimed better performance of modern varieties under optimal managementconditions but not under stress conditions (Sinebo 2005).

Conclusion

Apparently, low chickpea seed yield from early sowing in this environment resultsfrom low field emergence, high plant mortality, low plant stand at harvest, fewerpods per plant, fewer seeds per plant and low biomass production (Table 4). Finally,we conclude that appropriate sowing date, improved drainage methods andimproved varieties can substantially increase the productivity of chickpea onVertisols in Ethiopia.

References

Abebe M, Duffera M, Kidanu SY. 1994. Crops response to improved drainage of Vertisols inthe Ethiopian highlands. J Agron Crop Sci. 172:217–222.

112 G. Agegnehu and W. Sinebo

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2

Agegnehu G, Fikre A, Tadesse A. 2006. Cropping systems, soil fertility and crop managementresearch on cool-season food legumes in the central highlands of Ethiopia: a review. In: AliK, Keneni G, Ahmed S, Malhotra RS, Beniwal S, Makkouk K, editors. Food and foragelegumes of Ethiopia: progress and prospects. Proceedings of the Workshop on Food andForage Legumes, 2006, Addis Ababa, Ethiopia. Aleppo (Syria): CARDA. p. 135–145.

Asamenew G, Jutzi SC, Tedla A, McIntire J. 1988. Economic evaluation of improved Vertisoldrainage for food crop production in the Ethiopian highlands. In: Jutzi SC, Haque I,McIntire J, Stares JE, editors. Management of Vertisols in sub-Saharan Africa.Proceedings of a conference held at ILCA, Addis Ababa, Ethiopia, 31 August–4September 1987. Addis Ababa (Ethiopia): ILCA. p. 263–283.

Astatke A, Jabar M, Mohamed-Saleem MA, Erkossa T. 2002. Technical and economicperformance of animal-drawn implements for minimum tillage: experience on Vertisols inEthiopia. Exp Agric. 38:185–196.

Bejiga G, Tullu A, Tsegaye S. 1994. Effect of sowing date and seeding rate on the yield andother characters of chickpea (Cicer arietinum L.). Ethiopian J Agric Sci. 14:7–14.

Belayneh H. 1987. The effects of drainage system, drainage spacing and fertilizer on seed yieldand other characters of wheat, tef and chickpea on heavy clay soils of Ginchi. Ethiopian JAgric Sci. 8:85–94.

Ceccarelli S. 1994. Specific adaptation and breeding for marginal conditions. Euphytica.77:205–219.

Erkossa T, Stahr K, Gaiser T. 2006. Soil tillage and crop productivity on a Vertisol inEthiopian highlands. Soil Tillage Res. 85:200–211.

Eshete M. 1994. Chickpea and lentil agronomy research. In: Telaye A, Bejiga G, Saxena MC,Solh MB, editors. Cool-season food legumes of Ethiopia. Proceedings of the 1st NationalCool-Season Food Legumes Review Conference, 16–20 December 1993, Addis Ababa,Ethiopia. Aleppo (Syria): ICARDA. p. 230–251.

Haque I, Jutzi SC, Nnadi LA. 1988. Management of Vertisols for increased and stabilizedfood and feed production in Ethiopian highlands. In: Beyene D, editor. Soil scienceresearch in Ethiopia. Proceedings of the 1st Soil Science Research Review Workshop, 11–14 February 1986, Addis Ababa, Ethiopia. Addis Ababa (Ethiopia): IAR. p. 120–127.

Mamo T, Mohamed-Saleem MA. 2001. Soil and water research in Ethiopia. In: Dubale P,Dibabe A, Zeleke A, Ayele G, Kirub A, editors. Advances in Vertisols management in theEthiopian highlands. Proceedings of the International Symposium on Vertisol Manage-ment, 28 November–1 December 2000, Addis Ababa, Ethiopia. Debre Zeit (Ethiopia):EARO. p. 13–18.

SAS Institute. 2001. SAS/STAT user’s guide, version 8.2. Cary (NC): SAS Institute.Sinebo W. 2002. Yield relationships of barleys grown in a tropical highland environment.

Crop Sci. 42:428–437.Sinebo W. 2005. Trade off between yield increase and yield stability in three decades of barley

breeding in a tropical highland environment. Field Crop Res. 92:35–52.Wakeel AEl, Astatke A. 1996. Intensification of agriculture on Vertisols to minimize land

degradation in parts of the Ethiopian highlands. Land Degrad Dev. 7:57–67.

Archives of Agronomy and Soil Science 113

Dow

nloa

ded

by [

Mr

Get

ache

w A

gegn

ehu]

at 2

2:34

15

Febr

uary

201

2